VerSi. A method for the quantitative comparison of repository systems

Decision making and design processes for radioactive waste repositories are guided by safety goals that need to be achieved. In this context, the comparison of different disposal concepts can provide relevant support to better understand the performance of the repository systems. Such a task requires a method for a traceable comparison that is as objective as possible. We present a versatile method that allows for the comparison of different disposal concepts in potentially different host rocks. The condition for the method to work is that the repository systems are defined to a comparable level including designed repository structures, disposal concepts, and engineered and geological barriers which are all based on site-specific safety requirements. The method is primarily based on quantitative analyses and probabilistic model calculations regarding the long-term safety of the repository systems under consideration. The crucial evaluation criteria for the comparison are statistical key figures of indicators that characterize the radiotoxicity flux out of the so called containment-providing rock zone (einschlusswirksamer Gebirgsbereich). The key figures account for existing uncertainties with respect to the actual site properties, the safety relevant processes, and the potential future impact of external processes on the repository system, i.e., they include scenario-, process-, and parameter-uncertainties. The method (1) leads to an evaluation of the retention and containment capacity of the repository systems and its robustness with respect to existing uncertainties as well as to potential external influences; (2) specifies the procedures for the system analyses and the calculation of the statistical key figures as well as for the comparative interpretation of the key figures; and (3) also gives recommendations and sets benchmarks for the comparative assessment of the repository systems under consideration based on the key figures and additional qualitative arguments. We demonstrate the capabilities of the method by comparing two rather different repository systems: (1) a repository in salt with layout and site properties corresponding to the Gorleben site according to the state of knowledge and planning of 2007; (2) a fictive repository in the Lower Cretaceous claystones at a synthetic site in northern Germany with a layout corresponding to the repository planning in Switzerland and France in the years around 2005. The exemplary application of the method shows the kind of statements which can be derived and the restrictions that are often connected to these statements. In particular, it is shown that the two repository systems exhibit fundamentally different probability-consequence profiles. The main reason for this is that, according to the model calculations, radionuclide release from the salt repository occurs only in one scenario that does not represent the expected system evolution, whereas radionuclide release from the claystone repository occurs, though mostly at very low rates, in all scenarios. Such fundamentally different probability-consequence profiles have to be taken into account when evaluating the results of the presented or any other comparison method. It must be emphasized that no conclusions on the suitability of the considered host rocks in general or the Gorleben site in particular may be drawn from the exemplary results presented here.